Skip to content
Neato_Echo
Download

Manual Fix a problem

Local transcription isn't working

Parakeet or Whisper failing on your own PC — models, GPU acceleration, and the per-mode setting people miss.

Transcription fails or produces nothing, even though the model is downloaded.

Quick checks

  1. Confirm a model is downloaded

    Local transcription needs the model on disk. The model picker shows which are installed. If a download failed, or a model shows as downloaded but fails with "not downloaded", see A model won't download.

  2. Check the mode you're using has a model selected

    As above — the right tab, not just the right section.

  3. Try a smaller model

    Large models need substantial memory. If a small one works and a large one doesn't, that's the constraint, and the smaller model is a real answer rather than a workaround. Whisper Base is a quick test.

  4. Check bundled components weren't quarantined

    Security software removing whisper.cpp, sherpa-onnx or FFmpeg breaks local transcription specifically while leaving everything else working. See Antivirus blocks Neato Echo.

GPU acceleration

Whisper has an Enable GPU toggle. If GPU acceleration fails — an unsupported card, or not enough video memory — Neato Echo restarts on the CPU, retries the same request, and tells you it's using CPU instead. Your dictation still completes; it's slower.

If that happens every time, turn the GPU toggle off in the transcription model picker. It removes the failed attempt and the delay it adds.

What "local" does and doesn't need

  • No network once the model is downloaded. Nothing is sent anywhere.
  • Real CPU or GPU work — so a laptop on battery may throttle it, and modest hardware is slower.
  • The model selected for each mode, as above.

Local models covers the engines and what each needs from your PC.

FAQ

Transcription is slow.

Expected on modest hardware, especially with the larger Whisper models. A smaller model is the usual answer if speed matters more than the last few points of accuracy. For English dictation, Nemotron Speech Streaming EN is the fastest and shows text while you speak.

Which languages work locally?

It depends on the engine — the default Parakeet TDT 0.6B v3 and the Whisper models are multilingual; Nemotron Speech Streaming EN and Parakeet Unified EN are English-only. Local models has the detail.

Is local really private?

Yes. Transcription runs on your PC and the audio doesn't leave it. The only way audio goes elsewhere is if you add a key under Cloud Providers and pick that provider for a mode — and that's off by default.

Still need help?

Write to [email protected] with your Windows version, your Neato Echo version (Settings → System), and what you were doing. A person reads every message.

Email support Report a bug on GitHub